Ten years later (1707) Sir Robert Sibbald, using the materials collected. by Timothy Pont, Irvine, and David Buchanan, wrote as follows:

From thence [Shirva Burn] a large mile to Barhill, where was a great Fort, which hath had large Entrenchments, the ruins of Buildings were traced there, and many Stones have been found there with Inscriptions, and some with Figures upon them, which are kept at the Houses of the Nobility and Gentry in the Neighbourhood, there is a fresh Spring there and a Fountain, and amongst the Rubbish of the Fort, there was found a large Iron Shovel of a vast weight, and divers Sepulchres covered with large Stones, were found there upon digging the Ground.1

Sibbald's mention of the "fresh Spring" and the “ "Fountain is of interest. The latter is probably identical with the spring that still bubbles on the south side of the green basin. The former was in all likelihood the overflow from the buried well in the centre of the camp. If this surmise be correct, a further accumulation of debris on the surface must have almost completely choked the "Spring" soon after it was seen by Sibbald (or his authorities). There is no reference to it in the Itinerarium Septentrionale, and yet it is just one of the things that could hardly have failed to catch the eye of 'Sandy' Gordon, had it still been visible. His description is as follows:

:

[At Bar Hill there] is to be seen a very large and well preserved Fort upon the Wall: Here the Foundations of Buildings appear very distinct within the Area; which is surrounded with a considerable Number of Ditches and Ramparts, particularly at the East and West Ends of this Fort.... There is no Roman Fort, which I know of in Scotland, where the Vestiges of the old Buildings appear so plain as here, seeing the Prætorium, where the Præfect's Tent stood, is as yet very discernible, together with the Lodgements of the other Officers. . . . The military Way along Graham's Dike, divides itself into two Branches here, the one running by the side of the great Ditch, the other comes up to the Ramparts of this Fort.3

As it stands, the statement regarding the division of the road might be interpreted as perfectly accurate. Gordon's actual plan, however,

1 Historical Inquiries, p. 29.

4

2 The possible effects of mineral operations in the neighbourhood must also be reckoned with. It may be mentioned that the water of the well now rises to within 2 feet of the surface, at which level it stands.

3 Itin. Sept., pp. 54 f.

4

Op. cit., Plate 22.

is erroneous, and would appear to have been completed, not by the aid of observations on the spot, but by a literal interpretation of the text as printed, for the southern branch of the road, instead of soberly entering the fort by the eastern gate, is made to run full tilt against the ramparts.

Horsley, writing in 1732, was almost as much impressed by the remains as Gordon had been. He says:

Barhill fort deserves a particular regard and description. Its situation and strength, and the ruins of buildings within it are very remarkable. It has a triple rampart and a ditch on all sides but the north. The praetorium is visible, and of a similar figure within the fort itself. And three rows of ruins resembling ramparts and ditches appear within the praetorium. . . . There is a branch goes off from the principal military way to the north entry of this fort, and goes out again at the east entry, and then passing round the south side of the southern summit, comes up again to the main way.1

There is an obvious confusion here regarding the road, and the rampart is single, not triple. But the "three rows of ruins" (well shown, by the way, in Gordon's plan) were rediscovered during the recent excavations, when their true significance was made apparent. Maitland (1757) offers no fresh contribution of importance to our knowledge. As usual, his main anxiety is to detect flaws in the statements of Gordon and Horsley. Their accounts of the road give him an opening of which he takes full advantage. Unluckily, after he has administered a severe castigation to his predecessors for their stumbling, he himself falls headlong over precisely the same obstacle. "After the strictest search," he denies that the Military Way ran on the north of the fort. He is positive that it went straight through.2

Roy, in his Military Antiquities, deals very briefly with the Bar Hill station.

The fort, which is a little way detached from the south side of the wall, was probably one of those previously erected by Agricola. It is surrounded with double ramparts [and] contains many ruinous foundations within its area, whose vestiges, however, are not now so entire as represented in the Itinerarium.3

1 Britannia Romana, p. 169.

2 History of Scotland, pp. 176 f.

3 Op. cit., p. 160.

Roy, it will be seen, has been misled by surface appearances; as has already been remarked, the rampart is a single one. In his plan, too, he goes wrong about the roads just as Maitland had done, for he makes the Military Way traverse the camp from east to west. Yet his reference is exceedingly interesting for two reasons. He was the first observer to draw attention to the peculiarity presented by this fort in being completely detached from the body of the Vallum, a feature the true significance of which his military instinct enabled him to divine. Again, from what he says we can gather that the latter part of the eighteenth century saw many inroads on the ruins. One of these destructive raids seems to have taken place about 1790. In the old Statistical Account of Scotland1 (1791) the minister of Kirkintilloch, speaking of Bar Hill, says :

The fort is a square area of 150 yards. Some vaults belonging to it have lately been discovered. These are still entire; and are covered above with flat bricks, and floored with a mixture of lime and black and white gravel, resembling sand from the sea-shore, very unlike any that is now to be found in the neighbourhood.

During the early portion of the nineteenth century the process of quarrying went on apace. The site of the fort forms part of the estate of Gartshore, and in 1801 and 1802 the then proprietor carried out an extensive improvement scheme which was doubtless responsible for much. It may be to these changes that Stuart alludes when, writing in 1845, he tells us that

Many of [the foundations] have only been recently removed, to supply materials for building, or to serve the purpose of enclosing the adjacent fields.2

1 Vol. ii. p. 276.

2 Caledonia Romana (first ed.), p. 331. In the second edition, p. 338, a footnote (from another hand than Stuart's) gives a remarkable story of destruction said to have been wrought in 1809, when "stone walls" were "demolished" and "massive foundations rooted out." The accuracy of this whole statement is open to serious doubt. It is asserted, for instance, that the fort was surrounded by a thick stone wall forming a great square." Mr Whitelaw's excavations proved conclusively that this was not the case. The original narrator may have been confusing Bar Hill with Castlecary.

66

In 1892 the remains attracted the notice of the Glasgow Archaeological Society's Committee, then engaged on an examination of the structure of the Antonine Vallum. In their published Report they say :—

The outline of the station can still be made out in the field- the indent of the ditch all round being readily traceable, as well as the rounded corners of the enclosure.1

It might have been added that beneath the field hedge on the south the kerb of the southern rampart peeped out here and there above the grass. Even so, the picture presents a melancholy contrast to that drawn a century and a half before by Gordon. A few years longer, and the very site would perhaps have been forgotten. Fortunately, it was not to be so.

II. MR WHITELAW'S EXCAVATIONS.

An entirely fresh chapter in the history of the fort was opened in 1902. In the preface to the Report already quoted, cordial acknowledgment is made of the liberality with which Mr Alexander Whitelaw of Gartshore had placed at the service of the Glasgow Committee the labour necessary for cutting the numerous sections of Rampart and Ditch made at Croy and at Bar Hill. If his generosity deserved warm recognition then, Mr Whitelaw has now laid under a much deeper obligation all who are in any way interested in the story of Roman Britain. With a public spirit that is beyond praise, he has had the camp and its surroundings systematically explored at his own expense, keeping in close personal touch with the work throughout, and letting it be clearly understood that excavation was to proceed until there was nothing more to be discovered. It is but fair to add that the success achieved is due in no small measure to the enthusiasm, care, and well-reasoned perception of Mr John M'Intosh, the forester on the Gartshore Estate, to whom was entrusted the duty of immediate supervision. Mr M'Intosh has also rendered valuable aid in the preparation of the present Report.

Operations were commenced on November 20th, 1902. Attention 1 The Antonine Wall, etc., p. 94.

was first directed to those points where the surface indications were at all abnormal. The field had been under corn, and several patches showed stubble of unusually vigorous growth. An hour or two sufficed to dispose of these. On their being 'pitted,' the evidence was such as to suggest that, at some time or other, at least some of them had been fireplaces. Underneath each was a layer of wood ashes, from 1 to 2 feet thick, with a large stone in the centre. The spot next chosen for attack lay almost exactly in the middle of the fort. It had long been remarkable for its peculiar greenness in spring and early summer. The sloping ground immediately to the south of it, too, was frequently damp. Digging soon revealed the cause of these phenomena. Less than a foot beneath the surface the workmen struck the kerb of an old well.

Such a discovery on the very first morning was a piece of rare good fortune, and it was followed up without delay. In the face of considerable difficulties, the well (which had plainly been filled up of set purpose) was entirely cleared out. The upper stratum was disappointing. It consisted wholly of building material-pieces of freestone of various sizes, sometimes dressed, but generally quite rough--piled in hopeless confusion. At a depth of 12 feet there came to light the capital of a column, the precursor of much that was interesting. On November 22nd the workmen were 17 feet down, and had recovered five capitals and bases of pillars, 15 linear feet of round columns, and one fragment of an inscribed tablet. At this juncture it became necessary to erect overhead gear. In view of the great weight of the stones and the consequent danger of serious accident, it was deemed advisable to employ two winches, one to let down and pull up the man who attached the tackle for haulage, the other to bring to the surface the columns, bases, and capitals that now formed an almost solid mass, the larger pieces usually jammed hard against the stone 'cradling' of the well. With a total diameter of not more than 4 feet, the space conditions were extremely trying. The water also proved very troublesome, rising with steadily increasing rapidity. To keep it under, a running gear with two buckets had to be constantly in motion.